Comprehensive Association Analysis of 30 Single Nucleotide Polymorphisms Related with Metabolic Syndrome on Cancer Susceptibility in Japanese Population: A Case – Control Study

Cuneyd Parlayan1*, Shinobu Ikeda2, Noriko Sato2, Motoji Sawabe3, Masaaki Muramatsu2 and Tomio Arai4

1Department of Biomedical Engineering, School of Engineering and Natural Science & Regenerative and Restorative Medicine Research Center (REMER), Istanbul Medipol University, Beykoz, İstanbul/Turkey

2Department of Molecular Epidemiology, Medical Research Institute, Tokyo Medical and Dental University, Tokyo, Japan

3Department of Molecular Pathology, Graduate School of Health Care Sciences, Tokyo Medical and Dental University, Tokyo, Japan

4Department of Pathology, Tokyo Metropolitan Geriatric Hospital, Tokyo, Japan

*Corresponding Author:
Cuneyd Parlayan
Department of Biomedical Engineering
School of Engineering and Natural Science & Regenerative and Restorative Medicine Research Center (REMER)
Istanbul Medipol University
Beykoz, 34810, İstanbul/Turkey
Tel: +905312613827, 5312613827
Fax: 5312613827

Received Date: 15 January 2018; Accepted Date: 23 January 2018; Published Date: 30 January 2018

Citation: Parlayan C, Ikeda S, Sato N, Sawabe M, Muramatsu M, et al. (2018) Comprehensive Association Analysis of 30 Single Nucleotide Polymorphisms Related with Metabolic Syndrome on Cancer Susceptibility in Japanese Population: A Case - Control Study. Health Sci J. Vol. 12 No. 1: 545.

Copyright: © 2018 Parlayan C, et al. This is an open-access article distributed under the terms of the creative Commons attribution License, which permits unrestricted use, distribution and reproduction in any medium, provided the original author and source are credited.

Visit for more related articles at Health Science Journal


Single nucleotide polymorphisms (SNPs) related with metabolic syndrome may influence cancer. Our study aimed to replicate association studies of 30 selected polymorphisms pertaining to 21 genes and cancer phenotypes in Japanese elderly population. All polymorphisms were determined by genotyping on the samples collected from 1503 consecutive autopsy cases registered in the Japanese SNPs for geriatric research (JGSNP) data base. Current study consisted of 807 (54%) males and 696 (46%) females with a mean age of 80.3 ± 8.9 years. Number of males with cancer was 520 (57%) and females 398 (43%) in total of 918 (61%) cancer present cases and 585 (39%) controls without cancer. Our pathological cancer phenotypes consisted of gastric cancer (n=178), lung cancer (n=163), colorectal (n=137) and others. Our results indicated that rs1800470 (P10L) located TGFβI gene has significant association with overall cancer (CT: CC adjusted odds ratio (OR) 95% confidence interval (95% CI)=1.386 (1.017 – 1.888), p=0.007, TT: CC adjusted OR (95% CI)=1.438 (1.106 – 1.870), p=0.039, CT + TT: CC, adjusted OR (95% CI)=1.421 (1.111 – 1.816), p=0.005), and also with lung cancer ( CT + TT: CC adjusted OR (95% CI)=1.562 (1.024 – 2.383), p=0.039) in all subjects. We also found rs2243250 located in IL4 gene is associated with elevated risk of acute leukemia (TT + CT: CC adjusted OR (95% CI)=2.286 (1.243 – 4.206), p = 0.008). Furthermore, rs2077647 on ESRI gene was associated with decreased risk of acute leukemia (GG + AG: AA, adjusted OR (95% CI)=0.516 (0.319 – 0.835), p=0.007). We found that rs1800470, rs2243250 and rs2077647 polymorphisms may influence cancer in Japanese elderly population. Our study warrants further confirmations with larger samples.


Single nucleotide polymorphisms; Cancer risk; Acute leukemia; Lung cancer; Protective allele; TGFβ1; IL4; ESRI


Cancer is the most feared and considerably complex malady made of multifaceted combinations of genetic, epigenetic and environmental oncogenic risk factors. According to World Health Organization’s (WHO) estimation, there will be more than 13 million individuals affected by cancer in 2030 [1]. The recent studies clearly showed that Metabolic Syndrome (MetS) can be a major player creating carcinogenic risks [2-4]. Epidemiological studies indicated that diabetes obesity, dyslipidemia, hyperglycemia, inflammation, hyeperinsulinemia and other metabolic syndrome related diseases lead to cancer development [5,6]. Furthermore, there are number of evidences that specific cancer phenotypes including gastric cancer, prostate cancer, hepatocellular carcinoma, colorectal cancer, pancreatic cancer, and others demonstrated strong associations with metabolic syndrome [7-13].

Single nucleotide polymorphisms (SNPs) are the most common genetic modifications found in human genome. Recent genome wide association studies (GWAS) and candidate gene approach studies both indicated these genetic alterations, SNPs, influence cancer development [14,15]. The genetic modifications on metabolic syndrome related genes such as rs2243250 on Interleukin type 4 (IL4), rs2077647 on estrogen receptor 1 (ESR1) and rs1800470 (P10L) found in transforming growth factor beta type I (TGFβI) genes have been reported to influence cancer [16-18]. The rs2243250 (C-524T) located in the IL4 gene has been reported to be associated with diabetes [19]. The other polymorphism rs2077647 is associated with obesity, and also rs1800470 (P10L) polymorphism located in the intronic region causing Proline (CCG) to Leucine (CTG) functional change at codon 10 reported to have an association with metabolic syndrome [20,21]. Both IL4 and TGFβI genes possess both pro/antiinflammatory cytokines properties which are known to influence cancer development [22,23]. The rs1800470 polymorphism has been reported to effect the TGFβI gene expression level, and found to be associated with higher TGFβI serum level triggering the angiogenesis which often leads to cancer [24,25]. The rs2243250 (C-524T) polymorphism is located in the promoter region of IL4 which is a major cytokine involved in naive CD4 polarization, consequently binds to its receptor (IL4R) to cascade cytokine signalling that is a key part of the immune regulation, therefore aberrant expression of IL4 may affect the serum level of IL4 cytokine leading the cancer development [26-28]. On the other hand, rs2077647 (S10S) located on ESR1 gene causes synonymous codon alteration of serine (TCT) to serine (TCC) at the codon 10 position. Estrogen receptor 1 (ESR1) gene is not only involved in hormonal regulation, diabetes, insulin sensitivity and hypertension, but also endometrial carcinoma, cervical cancer, breast cancer, prostate cancer, and non-small cell lung cancer [29-37]. Moreover, rs1800470 (C29T) polymorphism interrupts the TGFβ signalling pathway disabling the tumor suppressing function, thus leads to oncogenesis including colorectal cancer, lung cancer, prostate pancreatic cancer, liver cancer and others [38-48]. Despite the numerous research efforts, the epidemiological association between metabolic syndrome and cancer development is still unclear and requires more scientific research.

Our objective was to investigate the metabolic syndrome associated SNPs illustrating clinical significance, whether they have influence on cancer in Japanese elderly population. In our study we examined 30 SNPs occurring on 21 different genes with varying functions including cytokines, enzymes, receptors and others that are related to metabolic syndrome but also have links or associations with cancer development. We focused on three SNPs, rs2243250 (C-524T), rs2077647 (S10S) and rs1800470 (P10L) which showed significant association(s) with overall cancer and specific cancer phenotypes.

Materials and Methods

1503 consecutive autopsy cases were analyzed, and mean age was 80.3 ± 8.9 years for all subjects. Our study included 807 males and 696 females. “C (+): cancer present” group is defined as any patient with at least one type of malignancy, and “C (-): cancer absent” group did not possess any type of malignancies. Cancer phenotypes were registered in the internet database of Japanese single-nucleotide polymorphisms (SNPs) for geriatric research (JG-SNP) were matched with our genotyping results for all 30 SNPs [49].

The autopsies were performed on 40% of patients who died in the Tokyo Metropolitan Geriatric Hospital between 1995 and 2004. The study subjects had high resemblance with the figures of the death causes reported by National Cancer Center, Japan (Vital Statistics of Japan) ( and previously reported, therefore our autopsy samples could be used [50,51]. All samples were pathologically reviewed and verified. The pathological assessments and genotyping were done in different institutions in double-blind fashion. Clinical data such as, smoking and drinking status, were retrieved from the patient’s medical records. Subjects were classified as “smokers” versus “non-smokers” by their smoking habit history, and also “drinker” versus “non-drinker” by the alcohol consumption history of subjects [52-60].

Written informed consent was obtained from the bereaved family of each of the patients prior to the autopsy examination. The use of autopsy materials for medical education and research was permitted by the Act of Postmortem Examinations of Japan. This study was approved by the Tokyo Medical and Dental University Ethics Committee (approval No. 2009-19-4) and the Tokyo Metropolitan Geriatric Hospital Ethics Committee (approval No.230405).


Genomic DNA was extracted from renal cortex of subjects by conventional procedures. Genotyping was done by TaqMan assay (Applied BioSystems) according to the protocols described by the manufacturers. PCR was performed in 5 μl reaction mix including 10 ng genomic DNA. PCR was conducted with the Light Cycler 480 instrument (Roche Diagnostics). Allelic discrimination of each polymorphisms were determined with ≥ 98% accuracy criteria. The genotyping results were determined Light Cycler Genotyping software (Roche Diagnostics). Genotyping accuracy was verified by randomly selected samples [61-65].

Statistical analysis

Statistical analysis was done by using SPSS software (IBM) version 19. The p-value (p) lower than 0.05 (p<0.05) was considered as statistically significant. Statistical significance in categorical values such as allelic distribution was calculated by 2-sided Fisher’s exact test. Bonferroni correction for multiple testing was not applied in our statistical analysis for its conservative nature. Continuous values such as age was analyzed by analysis of variance (ANOVA). Hardy-Weinberg equilibrium (HWE) was assessed using permutation test. Chisquare test was used to determine significance in distribution of genotypes in cancer present or absent subjects. Odds ratio (OR), 95% Confidence interval (95% CI) values were calculated by binary logistic regression analysis. Confounding parameters such as “gender, smoking and drinking status, and age” were adjusted in regression analysis. In genetic models we used following genotypic groupings; additive model (AA, Aa, aa) “A” is the dominant allele and “a” is the recessive allele. Dominant model was defined as (AA vs. Aa + aa), recessive model (aa vs. Aa + AA) and co-dominant model (AA vs. aa) [66-70].


General characteristics of subjects

Demographics of our study are shown in Table 1. Our current study consisted of 1503 consecutive autopsy subjects including 918 (61%) cancer present and 585 (39%) cancer absent control subjects. Mean age was 80.3 ± 8.9 and also significantly different between cancer present and absent subjects (p<0.001). Males possessing cancer were 520 (57%) and females 398 (43%), gender parameter was found to be statistically significant between cancer present and absent subjects (p=0.004). Smokers, 460 (50%), were higher than nonsmokers, 408 (44%) in cancer present subjects. Drinker subjects, 324 (35%) were less than non-drinker subjects, 533 (58%) in cancer bearing subjects. No statistical significance was found for both smoking (p=0.134) and drinking status (p=0.075). Cancer bearing subjects were further stratified; single cancer bearers were 689 (46%), and more than two cancer types’ bearers were 229 (15%). Our pathologically verified cancer phenotypes included gastric cancer (n=178), lung cancer (n=163), colorectal (n=137) and others (Table 1).

Table 1 Demographic variables of examined subjects.

Characteristics Total, n (%)
(n = 1503)
C (-)¹, n (%)
(n = 585)
C (+)², n (%)
(n = 918)
Age at death, (years) 80.3 ± 8.9 81.3 ± 9.2 79.6 ± 8.6 <0.001
Gender, n (%)
Male 807 (54) 287 (49) 520 (57) 0.004
Female 696 (46) 298 (51) 398 (43)
Smoking status, n (%)
Smoker 717 (48) 257 (44) 460 (50)  0.134
Non-smoker 677 (45) 269 (46) 408 (44)
Missing 109 (7) 59 (10) 50 (6)
Alcohol consumption, n (%)
Drinker 499 (33) 175 (30) 324 (35) 0.075
Non-drinker 887 (59) 354 (60) 533 (58)
Missing 117 (8) 56 (10) 61 (7)
Cancer, n (%)
0   585 (39)    
1     689 (46)  
≥2     229 (15)  
Cancer sites3, n
Gastric     178  
Lung     163  
Colorectal     137  
Prostate     94  
Acute Leukemia     78  

Characteristics of examined polymorphisms

The general characteristics of polymorphisms are shown in Table 2. We analyzed 30 SNPs formerly found to have associations with metabolic syndrome or metabolic diseases including diabetes, myocardial infraction, arterial calcifications, cholestasis and others. Among these 30 polymorphisms, rs2077547, rs2243250 and rs1800470 SNPs illustrated significant results, thus we further focused on these three polymorphisms (Table 2). The rs2077647 (A/G) polymorphism located in the 6th chromosome which causes synonymous change of Serine (S) to Serine (S) in the estrogen receptor 1 (ESR1) gene. The ESR1 (Gene ID: 2099) receptor / enzyme gene has clinical significance mainly in HDL metabolism and atherosclerosis, and is registered in Online Mendelian Inheritance in Man (OMIM, registry number: 133430). ESR1 has been reported to have association with breast cancer, endometrial cancer and involved in sexual development. The other SNP rs2243250 (C/T) polymorphism is located in the 5th chromosome and associated with Interleukin 4 (IL4) cytokine (Gene ID: 3565). IL4 is produced by activated T cells and key functions in immune system. IL4 possesses clinical significance (OMIM registry number: 147780). The rs1800470 is located on transforming factor beta type 1 (TGFβI) gene (Gene ID: 7040). In rs1800470 polymorphism C to T nucleotide modification causes Proline (P) codon change to Leucine (L) at the 10th codon (P10L). TGFβI gene has also clinical importance (OMIM registry number: 190180) and reported to have involvements in obesity and cancer [71-80].

Table 2 General characteristics of rs2077647, rs2243250 and rs1800470 polymorphisms.

SNP Allelic Change Strand Location (Ch:Loc)1 Target Gene Gene ID2 Functional change Functions of genes Key features of genes2 OMIM3 MetS.4 Can.5
rs2077647 A/G Reverse strand 6:152129077 Estrogen receptor (ESRI) 2099 (S10S)* Receptor/Enzyme Estrogen receptor, breast cancer, endometrial cancer, sexual development 133430 70 71
rs2243250 C/T Forward strand 5:132009154 Interleukin 4 (IL4) 3565   Cytokine/ hormone / growth factor Is a cytokine produced by activated T cells. Involved in immune regulatory signalling 147780 72 73
rs1800470 C/T Reverse strand 19:41858921 Transforming growth factor, beta type I (TGFβI) 7040  (P10L)* Cytokine/ hormone / growth factor Involved in proliferation, differentiation, adhesion, migration and other types of cellular functions. 190180 74 75

Allelic and genotypic distributions of examined polymorphisms

The allelic and genotypic distributions of selected polymorphisms, rs2077647, rs2243250 and rs1800470, in all subjects are shown in Table 3, and the results of the entire polymorphisms. The allelic distribution of rs2077647’s A: G was 1733 (58%): 1251 (42%), genotypic distribution of AA: AG: GG was 516 (35%): 701 (47%): 275 (18%) and observed minor allele frequency was 0.42. The allelic distribution of rs2243230’s T: C was 2036 (68%): 948 (32%) genotypic distribution of TT: CT: CC was 696 (47%): 644 (43%): 152 (10%) and observed minor allele frequency was 0.32. Finally, the allelic distribution of rs1800470’s C: T was 1535 (52%): 1429 (48%), genotypic distribution of CC: CT: TT was 401 (27%): 733 (50%): 348 (23%) and observed minor allele frequency was 0.48. Due to stringent genotyping criteria some samples were either excluded or not successfully genotyped, these missing samples were as following; rs2077647 (n=11), rs2243250 (n=11) and rs1800470 (n=21). All of these three polymorphisms were consistent with Hardy-Weinberg Equilibrium (p=0.174, p=0.866 and p=0.713) respectively (Table 3). All of our genotyped polymorphisms did not show any deviation from Hardy-Weinberg Equilibrium.

Table 3 Allelic and genotypic distribution of rs2077647, rs2243250 and rs1800470 polymorphisms in all subjects.

SNP Alleles Genotypes p(HWE)b MAF
A G AA AG GG    
rs2077647 1733 1251 516 701 275    
(%)a -58 -42 -35 -47 -18 0.174 0.42
Missing 11            
T C TT CT CC    
rs2243250 2036 948 696 644 152    
(%) -68 -32 -47 -43 -10 0.866 0.32
Missing 11            
C T CC CT TT    
rs1800470 1535 1429 401 733 348    
(%) -52 -48 -27 -50 -23 0.713 0.48
Missing 21            

Genotypic distribution of examined polymorphisms in different inheritance models

We analyzed all SNPs with additive, dominant, recessive and co-dominant genetic models to find their relative effect in overall cancer in all subjects. The results of rs2077647, rs2243250 and rs1800470 is shown in Table 4, and the rest of the results. The rs2077647 polymorphism found in ESR1 gene showed significantly high frequency of AA genotype in cancer present subjects [81-86]. The additive model AA: AG: GG genotypic distribution in cancer present subjects was 335 (37%): 412 (45%): 167 (18%) and did not show significance (p=0.092). Dominant model AA: GG + AG genotypic distribution in cancer bearing subjects was 335 (37%): 579 (63%), and demonstrated statistical significance (p=0.039). Recessive model GG: AA + AG and co - dominant model AA: GG genotypic distributions in cancer present subjects was 167 (18%): 747 (82%) and 335 (67%): 167 (33%). Both recessive and co-dominant models did not show statistical significance between cancer present and absent subjects (p=0.837 and p=0.246) respectively (Table 4).

Table 4 Genotypic distribution of rs2077647, rs2243250 and rs1800470 polymorphisms in four different models in overall cancer.

SNP(Gene) Additive C -an (%) C +an (%) Dominant C -n (%) C +n (%) Recessive C -n (%) C +n (%) Co-dominant C -n (%) C +n (%)
AA 181 (31) 335 (37) AA 181 (31) 335 (37) GG 108 (19) 167 (18) AA 181 (63) 335 (67)
AG 289 (50) 412 (45)                  
GG 108 (19) 167 (18) AG + GG 397 (68) 579 (63) AG + AA 470 (81) 747 (82) GG 108 (37) 167 (33)
    0.092     0.039     0.837     0.246
TT 272 (47) 424 (47) TT 272 (47) 424 (47) CC 47 (8) 105 (12) TT 272 (85) 424 (80)
CT 261 (45) 383 (42)                  
CC 47 (8) 105 (11) CT + CC 308 (53) 488 (53) CT + TT 533 (92) 807 (88) CC 47 (15) 105 (20)
    0.091     0.915     0.035     0.065
  CC 177 (31) 224 (25) CC 177 (31) 224 (25) TT 126 (22) 222 (24) CC 177 (58) 224 (50)
rs1800470 CT 273 (48) 460 (51)                  
(TGFβI) TT 126 (21) 222 (24) CT + TT 399 (69) 682 (75) CT + CC 450 (78) 684 (76) TT 126 (42) 222 (50)
pb     0.038     0.012     0.258     0.03

Genotypic distributions of rs2243250 located in IL4 gene in the cancer present subjects were as following; additive model TT: CT: CC was 424 (47%): 383 (42%): 105 (11%) and was not found to be significant (p=0.091), dominant model TT: CT + CC was 424 (47%): 488 (53%) and was not significant (p=0.915), recessive model CC: CT + TT was 105 (12%): 807 (88%) and CC genotype was significantly high in cancer present subjects (p=0.035), co-dominant model TT: CC was 424 (80%): 105 (20%) and was not significant (p=0.065), (Table 4).

The rs1800470 located in TGFβI gene polymorphism’s genotypic distributions of four different models in cancer bearing subjects were as following; additive CC: CT: TT was 224 (25%): 460 (51%): 222 (24%) and TT genotype was significantly higher in cancer present subjects (p=0.038), dominant model CC: CT + TT was 224 (25%): 682 (75%) and T allele was significantly high in cancer present subjects (p=0.012), recessive model TT: CT + CC was 222 (24%): 684 (76%) and no statistical significance was observed (p=0.258), co-dominant model CC: TT was 224 (50%): 222 (50%) and TT genotype was found to be statistically higher in cancer bearing subjects (p=0.030), (Table 4).

Association analysis of rs2077647, rs2243250 and rs1800470 polymorphisms with overall cancer

To find out the possible association(s) of overall cancer risk with rs2077647, rs2243250 and rs1800470 polymorphisms, logistic regression analysis was applied and the results is shown in Table 5. The rs2077647 polymorphism case, AG, GG and GG + AG genotypes were tested with respect to major genotype AA, there were statistical significance in crude analysis AG: AA (p=0.031) and GG + AG: AA (p=0.032), however after adjustment with age, gender, smoking and drinking status, all examined genotypes including the GG + AG combination lost their statistical significance; AG: AA (p=0.051), GG: AA (p=0.261), GG + AG: AA (p=0.055). Therefore, there was no statistically reliable association between rs2077467 and overall cancer risk [87-93].

Table 5 Association analysis of rs2077647, rs2243250 and rs1800470 polymorphisms with overall cancer in all subjects.

SNP(Gene) Genotype C -an (%) C +a n (%) Crude OR (95% CI) p* Adjusted ORb (95% CI) p*
AA 181 (31) 335 (37) 1.000 (reference)   1.000 (reference)  
AG 289 (50) 412 (45) 0.761 (0.594 - 0.975) 0.031 0.780 (0.607 - 1.001) 0.051
GG 108 (19) 167 (18) 0.812 (0.591 - 1.115) 0.198 0.832 (0.604 -1.147) 0.261
AG + GG 397 (68) 579 (63) 0.775 (0.614 – 0.978) 0.032 0.794 (0.627 - 1.005) 0.055
rs2243250(IL4) TT 272 (47) 424 (47) 1.000 (reference)   1.000 (reference)  
CT 261 (45) 383 (42) 1.393 (0.941 - 2. 062) 0.098 1.385 (0.932 - 2.058) 0.107
CC 47 (8) 105 (11) 0.915 (0.727 - 1.152) 0.452 0.915 (0.725 - 1.155) 0.454
CT + TT 533 (92) 807 (88) 1.000 (reference)   1.000 (reference)  
CC 47 (8) 105 (12) 1.451 (0.998 - 2.118) 0.051 1.446 (0.989 - 2. 113) 0.057
CC 177 (31) 224 (25) 1.000 (reference)   1.000 (reference)  
CT 273 (48) 460 (51) 1.402 (1.081 - 1.818) 0.038 1.386 (1.017- 1.888) 0.007
TT 126 (21) 222 (24) 1.383 (1.018 - 1.879) 0.011  1.438 (1.106 - 1.870) 0.039
CT + TT 399 (69) 682 (75) 1.396 (1.094 - 1.780) 0.007 1.421 (1.111-1.816) 0.005

In rs2243250 polymorphism CC and CT was tested against to major genotype TT. There was no significant results in both in crude CT: TT (p=0.098) and CC: TT (p=0.452) and adjusted regression analysis CT: TT (p=0.107) and CC: TT (p=454). Furthermore, CC: CT + TT was not significant in crude (p=0.051) and adjusted (p=0.057) analysis. Therefore, no statistical association was found between rs2243250 polymorphism and overall cancer risk [94-97].

The rs1800470 showed following results in the comparison of CT, TT and CT + TT with respect to major genotype CC. All three genotypic comparisons were statistically significant in crude results; CT: CC OR (95% CI)=1.402 (1.081 – 1.818), p=0.038. TT: CC OR (95% CI)=1.383 (1.018 – 1.879), p=0.011. CT + TT: CC OR (95% CI)=1.396 (1.094 – 1.780), p=0.007. In adjusted results the significant associations were also kept. CT: CC adjusted OR (95% CI)=1.386 (1.017 – 1.888), p=0.007. TT: CC adjusted OR (95% CI)=1.438 (1.106 – 1.870), p=0.039 [98-100].

CT + TT: CC adjusted OR (95% CI)=1.421 (1.111 – 1.816), p=0.005. Thus, rs1800470 polymorphism T allele is associated with 42% higher overall cancer risk in our samples.

Genotypic distribution of rs2077647, rs2243250 and rs1800470 polymorphisms in specific cancer phenotypes

We continued to investigate the genotypic frequencies in specific cancer phenotypes consisting of gastric cancer, lung cancer, colorectal cancer and acute leukemia in additive, dominant, recessive and co-dominant genetic models for all subjects and the results are shown in Table 6. Both rs2077647 and rs22432250 polymorphisms did not show any statistical significance in the specific genotypic distribution in each additive, dominant, recessive and co-dominant genetic models for gastric cancer, lung cancer and colorectal cancer (Table 6). On the other hand, both rs2077647 and 22432250 polymorphisms had significant results in acute leukemia. The AA genotype frequencies in rs2077647 polymorphism were 37 (7%) in additive, recessive and co-dominant models and was found to be significantly higher in acute leukemia (p=0.023, p=0.014 and p=0.015 respectively) in all subjects. Additionally, CC genotype of rs2243250 polymorphism was 16 (10%) in additive, dominant and co-dominant models and had significantly high frequency (p=0.008, p=0.006 and p=0.010 respectively) in all subjects.

Table 6 Genotypic distribution of rs2077647, rs2243250 and rs1800470 polymorphisms in four different models in specific cancers.

SNP (Gene) Cancer type Additive Dominant Recessive Co-dominant
AA AG GG pa AA AG + GG pb GG AG + AA pb AA GG pb
rs2077647 (ESRI) Gastric 67 (13) 81 (12) 29 (11) 0.567 67 (13) 110 (11) 0.355 29 (11) 148 (12) 0.536 67 (13) 29 (11) 0.361
Lung 60 (12) 67 (10) 36 (13) 0.241 60 (12) 103 (11) 0.542 36 (13) 127 (11) 0.201 60 (12) 36 (13) 0.568
Colorectal 47 (9) 63 (9) 26 (10) 0.976 47 (9) 89 (9) 1 26 (10) 110 (9) 0.817 47 (9) 26 (10) 0.898
Acute Leukemia 37 (7) 32 (5) 8 (3) 0.023 37 (7) 40 (4) 0.014 8 (3) 69 (6) 0.069 37 (7) 8 (3) 0.015
TT CT CC   TT CT + TT   CC CT + TT   TT CC  
Gastric 81 (12) 85 (13) 12 (8) 0.184 81 (12) 97 (12) 0.81 12 (8) 166 (12) 0.114 81 (12) 12 (8) 0.2
rs2243250 (IL4) Lung 72 (10) 79 (12) 12 (8) 0.244 72 (10) 91 (11) 0.561 12 (8) 151 (11) 0.271 72 (10) 12 (8) 0.454
Colorectal 62 (9) 53 (8) 19 (12) 0.253 62 (9) 72 (9) 1 19 (12) 115 (9) 0.132 62 (9) 19 (12) 0.173
Acute Leukemia 32 (5) 30 (5) 16 (10) 0.008 32 (5) 46 (6) 0.352 16 (10) 62 (5) 0.006 32 (5) 16 (10) 0.01
  CC CT TT   CC CT + TT   TT CT + CC   CC TT  
Gastric 43 (11) 82 (11) 51 (15) 0.183 43 (11) 133 (12) 0.418 51 (15) 125 (11) 0.072 43 (11) 51 (15) 0.122
rs1800470 (TGFβI) Lung 32 (8) 97 (13) 32 (9) 0.013 32 (8) 129 (12) 0.031 32 (9) 129 (11) 0.279 32 (8) 32 (9) 0.601
Colorectal 36 (9) 68 (9) 30 (9) 0.936 36 (9) 98 (9) 1 30 (9) 104 (9) 0.831 36 (9) 30 (9) 0.898
Acute Leukemia 16 (4) 35 (5) 24 (7) 0.169 16 (4) 59 (6) 0.287 24 (7) 51 (5) 0.092 16 (7) 24 (7) 0.102

Although there were significant results in both rs2077467 and rs2243250 polymorphisms, one cautiously ought to consider the low sample number of acute leukemia (n=78) (Table 7).

Table 7 Association analysis of rs2077647, rs2243250 and rs1800470 polymorphisms with acute leukemia and lung cancer in all subjects.

SNP (Gene) Cancer type Genotype C +a n (%) Crude OR (95% CI) p* Adjusted ORb (95% CI) p*
    AA 37 (7) 1.000 (reference)   1.000 (reference)  
 rs2077647 (ESRI)  Acute Leukemia AG + GG 40 (4) 0.493 (0.307 -0.792) 0.003 0.516 (0.319 -0.835) 0.007
CT + TT 162 (5) 1.000 (reference)   1.000 (reference)  
 rs2243250 (IL4)  Acute Leukemia CC 6 (10) 2.322 (1.280 - 4.213) 0.006 2.286 (1.243 - 4.206) 0.008
CC 32 (8) 1.000 (reference)   1.000 (reference)  
 rs1800470 (TGFβI)  Lung CT + TT 129 (12) 1.560 (1.031 - 2.359) 0.035 1.562 (1.024 - 2.383) 0.039

Association analysis of rs2077647, rs2243250 and rs1800470 polymorphisms with specific cancer phenotypes

We further investigated the possible association(s) of the selected polymorphisms, rs2077647, rs2243250 and rs1800470, with specific cancer phenotypes, and significant results are shown in Table 7. All 30 SNPs were screened through for association(s) of specific cancer phenotypes (data not shown) however, only three SNPs illustrated significant association with acute leukemia and lung cancer (Table 7). The rs2077647 polymorphism located on ESRI indicated that AG + GG: AA comparison both in crude analysis OR (95% CI)=0.493 (0.307 – 0.792) and adjusted analysis Adjusted OR (95% CI)=0.516 (0. 319–0. 835) found to be statically associated (p=0.003 and p=0.007 respectively) with reduced acute leukemia. Thus, G allele had a protective feature and may reduce the acute leukemia risk up to 48% in all subjects (Table 7).

The rs2243250 polymorphism located in IL4 showed that CC: CT + TT ratio both in crude analysis OR (95% CI)=2.322 91.280 – 4.213) and adjusted analysis adjusted OR (95% CI)=2.286 (1.243 – 4.206) was found to be significantly associated (p=0.006 and p=0.008 respectively) with the increased risk of acute leukemia (Table 7). Although, CC genotype significantly increased the risk of acute leukemia, the low number of acute leukemia and its effect ought to be considered to make a concrete conclusion.

Finally, rs1800470 polymorphism located on TGFβI demonstrated that CT + TT: CC genotypic comparison both in crude analysis OR (95% CI)=1.560 (1.031 – 2.359) and adjusted analysis adjusted OR (95% CI)=1.562 (1.024–2.382) were significantly (p=0.035 and p=0.039 respectively) associated with increased lung cancer. As a result, T carriers increased 56% lung cancer occurrence in all subjects (Table 7).


In our current research, we hypothesized if the metabolic syndrome related SNPs may influence primarily overall cancer risk and secondarily the risk of specific cancer phenotype(s) in elderly Japanese population. Our study population consisted of 1503 individuals in which 61% had cancer, with mean age of over 80 years old (Table 1). 30 SNPs found in 21 encoding genes having clinical significance and associated with metabolic syndrome were analyzed (Table 2). Among all 30 genotyped SNPs, we focused on three polymorphisms, rs2077547, rs2243250, rs1800470 polymorphisms (Table 3) and examined in additive, dominant, recessive and codominant genetic models (Table 4). Then, these three polymorphisms were investigated for other specific cancer phenotypes including gastric cancer, lung cancer, colorectal cancer, and acute leukemia (Table 6), other cancer phenotypes such as liver, pancreas etc. were also analyzed but no significance was detected (data not shown). Finally, the association(s) between rs2077547, rs2243250, and rs1800470 polymorphisms and overall cancer and selected specific cancer phenotypes including acute leukemia and lung cancers were analyzed (Tables 5 and 7).

We found that G allele in rs2077647 (S10S) polymorphism had significantly higher frequency in control group in dominant model (Table 4). Furthermore, there was no association between rs2077647 and overall cancer risk (Table 5). Our findings indicated that G allele was significantly higher in control subjects in both additive and dominant model for acute leukemia (Table 6). As we further investigated the relationship between rs2077647 G allele carriers with acute leukemia, G allele appeared to be associated with reduced level of acute leukemia, thus had a “protective” effect against acute leukemia in all subjects (Table 7). Our findings are consistent with previous reports; Nicolaiew et al., Wang et al. and Chae et al. found no significant association between ESR1 related polymorphisms such as rs2077647 with cancer more specifically prostate cancer [100-102]. Fernandez et al. found similar result that rs2077647 (S10S) polymorphism had a protective function against breast cancer in Spanish population [103] and a similar report from a Taiwanese group, Hsiao et al. S10S variant showed protective property [104]. The underlying reason(s) of this observed protective property of G allele against acute leukemia is still unclear, however, rs2077647 may have unknown tumour suppressor effect either by itself or with other SNPs having high linkage disequilibrium or may interact with other genetics factors to supress the oncogenic formation. Also, there are some reports asserted that consumption of phytoestrogens or isoflavone for instance soy beans, a dominant dietary intake in Japan, has protective effect against androgenic pathway related cancer types such as prostate cancer [105]. Thereby, G carriers may initiate analogous pathway(s) to reduce the risk of tumour development.

The other important finding was regarding to rs2243250 (C-534T) polymorphism found on IL4 gene. C allele was minor allele (Table 3) and found significantly frequent in case subjects in recessive model (Table 4). However, no significant association was found with overall cancer risk (Table 5). As we further checked the possibilities of rs2243250 (C-534T) polymorphism for other specific cancer phenotypes, C allele appeared to be the risk allele in additive and dominant models of acute leukemia (Table 6), consequently the C carriers found to be significantly associated with acute leukemia (Table 7). Our results are aligned with the previous reports such, De Guia et al. found rs2243250 (C-534T) polymorphism, C carriers had higher serum level of IgE leading to higher chance of getting allergy, thus leading to cancer [106]. There are other reports indicated that T to C allelic alteration in rs2243250 may lead to aberrant T-helper 2 cells, then C or CC carriers have higher tendency to have cancers such oral, head and neck and leukemia [107,108]. Moreover, there are some conflicting results reported by other research groups indicating that T allele had significant oncogenic effect [27]. Additionally, Zhenzhen et al. found CC homozygote associated with increased renal cancer, but decreased oral cancer risk in Asian population [26]. The underlying reasons or mechanisms of the oncogenic effect of rs2243250 (C-534T) polymorphism remains largely unknown, however some possible explanations can be asserted; first, the aberrant T-helper type 2 cells may disturb the immune surveillance causing less defensive immune system against either cancer via allergies or directly cancer, more specifically leukemia or haematological malignancies [109]. Second, IL-4 is also a soluble pro-inflammatory cytokine, thus the dysfunctional cytokine pathway (s) due to polymorphic allele alteration may lead to immunosuppressive state coupling with tumour initiating factors, and thereby oncogenesis occurs [110]. Third, there may be ethnical differences as in the Zhenzhen et al. report in which Caucasians and Asian possess differential risk factors for different types of cancers caused by aberrant IL-4 gene [26].

Our final and most critical results are pertaining to rs1800470 (P10L) polymorphism. There is a functional change at the 10th codon on transforming growth factor beta type I (TGFβI) gene from proline (P) to leucine (L) amino acids located on chromosome 19. T allele was minor allele (Table 3). T allele and TT homozygote genotype showed significantly high presence in additive, dominant and co-dominant models in overall cancer risk (Table 4). Furthermore, T allele carriers had significant association with overall cancer risk in Japanese elderly population (Table 5). T allele was also found in high frequency in lung cancer case both in additive and dominant models (Table 6). Finally, the T carriers found to be associated with 56% higher lung cancer risk (Table 7). Our results indicated that 10L (Leucine) substitution had significant association with overall and lung cancer risk.

There are numerous reports about TGFβI, which is most abundant form of TGFβ gene family, associated polymorphisms and TGFβ signalling pathway and their oncogenic effects in the literature [23-25,40-48,111-122]. The rs1800470 polymorphism causing cysteine to thymine (C→T) at 29th nucleotide resulting the change of proline (CCG) amino acid to leucine (CTG) at 10th codon (P10L). T allele carriers were previously reported to have higher plasma concentration of TGFβI [42,111]. The Our results were consistent with previous reports such as Pooja et al. found T allele carriers had higher risk and significant association with breast cancer in Indian population [112]. Falleti et al. found TT genotype carriers had significantly higher in hepatocellular carcinoma [45]. Three Japanese groups also confirmed that T or TT carriers had high risk of developing prostate cancer, gastric cancer and hepatocellular carcinoma [113-115]. There are conflicting results, despite of the several meta-analysis using considerably higher number of samples [18]. Zhang et al. found in their meta-analysis that rs1800469 (-509°C/T), which is in considerably similar linkage disequilibrium (LD) block with rs1800470, is associated with upper digestive tract cancer but not rs1800470 polymorphism [116]. One study conducted in Spain asserted that pro-and anti-inflammatory cytokine gene polymorphisms is not linked with gastric cancer [117]. Other groups also indicated rs1800470 T carriers have either no association or association with decreased level of prostate, colorectal and gastric cancer [39,42,118]. It is still largely unknown why there are several conflicting reports, however, there are possible reasons that may be asserted; firstly, the ethnical back ground is a key factor of these different results [18]. Secondly, type of the study i.e., hospital based, multicenter based, population based etc. may increase the selection bias, heterogeneity of the sampling, testing errors, or false positives [119]. Thirdly, genetically related diseases some genotypes may have “low-penetrance” effect, thus either C carriers (Proline) or T carriers (Leucine) may have low penetrance which may create differing results in cancer phenotypes [120]. Fourthly, the unique dual character of TGFβI gene might have been the reason. TGFβI can either suppress or cause oncogenesis, thereby T carriers could have oncogenic property of TGFβI gene resulting overall cancer risk by creating aberrant TGFβ pathway signalling specifically effecting either SMAD protein or receptor (TGFβRI/II) binding [23,25,46,48].

Among other results pertaining to rs1804700 polymorphism on TGFβI gene, T carrier was also associated with elevated lung cancer (Table 7). Our finding was supported by two studies done in Asian population; both studies indicated T or TT carriers have elevated risk of lung cancer as compared with other genotypes or alleles [40,41]. Although the reasons behind lung cancer risk caused by rs1800470 is still world-wide undergoing research topic, there are possible reasons; firstly the genetic alteration may initiate the epithelial-mesenchymal transition (EMT) like growth which often leads to cancer, besides the lung tissues is known for the vulnerable nature for the invasive epithelial growth [121-123]. Secondly, canonical SMAD 2/3 then SMAD4 proteins binding(s) may have been hindered then it leads to either perturbed or loss of TGFβ signalling causing higher tendency of metastatic invasion or inhibition of tumor supressing function of TGFβ protein families enabling cancer development [23-25]. Finally, rs1800470 may interact with other unknown SNP(s) or gene(s) to create genetic or epigenetic oncogenic risk factor(s) coupled with the aberrant TGFβ signalling.

Another interesting findings were rs1042522 polymorphism located on Tumour protein p53 (TP53) gene involved in cell cycle arrest and DNA repair, and rs1799964 polymorphism occurring on Tumor Necrosis Factor (TNF) gene involved in proliferation and apoptosis, did not show any significant result, despite the fact, both genes and SNPs were well-known for their oncogenic affects [124,125]. Counter intuitively, genetic modifications on both TNF and TP53 genes did not show any significant association(s) or cancer risk in our current study.

Our study had some limitations; our samples were hospital based consecutive autopsy cases, this fact may cause selection bias due to chance of admission, age bias having elderly individuals, survival bias associated with other disease(s) causing cancer development in particular genotype. Our data lacked the cancer progression which may have provided some critical depth in our results. Furthermore, the life style information for instance exercising and dietary habits were not included in our data. There was also no data on hematological properties i.e., blood proteins, lipid concentration etc. which could provide important insights for better understanding of cancer development [51].


Our results indicated that rs2077647 (S10S) and rs2243250 (C-534T) polymorphisms did not show overall cancer risk, but rs2077647 may reduce and rs2243250 may increase acute leukemia risk. On the other hand, rs1800470 (P10L) polymorphism may pose overall cancer and lung cancer risk in Japanese elderly population. Although, these three SNPs showed significant results, the concrete conclusions cannot be made due to the low sample numbers. Our findings shed some light on the understanding of common genetic modifications such as rs2077647, rs2243250 and rs1800470 polymorphisms associated with metabolic syndrome may influence cancer development. Further studies accompanied with functional analysis with larger samples are warranted to verify our current results.

Authors’ Contributions

CP conducted genotyping analysis, statistical analysis, and drafted the manuscript. MS and TA provided pathological & clinical data, and also provided advice on the manuscript. SI assisted on data retrieval and gave genotyping analysis. NS provided critical advice. MM proofread, edited, and provided critical comments and guidance.


This study was supported by the Joint Usage/Research program of Medical Research Institute, Tokyo Medical and Dental University, and Smoking Research Foundation. We thank all Molecular Epidemiology laboratory members of Tokyo Medical and Dental University for their comments and support during our study.


Select your language of interest to view the total content in your interested language

Viewing options

Post your comment

Share This Article

Flyer image
journal indexing image

Post your comment

captcha   Reload  Can't read the image? click here to refresh